1 chapter 20 transaction management transparencies © pearson education limited 1995, 2005

1

Chapter 20

Transaction Management

Transparencies

© Pearson Education Limited 1995, 2005

2

Chapter 20 - Objectives

Function and importance of transactions. Properties of transactions. Concurrency Control

– Meaning of serializability.– How locking can ensure serializability.– Deadlock and how it can be resolved.– How timestamping can ensure

serializability.– Optimistic concurrency control.– Granularity of locking.


3

Chapter 20 - Objectives

Recovery Control– Some causes of database failure.– Purpose of transaction log file.– Purpose of checkpointing.– How to recover following database

failure. Alternative models for long duration transactions.


4

Transaction Support

Transaction

Action, or series of actions, carried out by user or application, which reads or updates contents of database.

Logical unit of work on the database. Application program is series of transactions with non-

database processing in between. Transforms database from one consistent state to

another, although consistency may be violated during transaction.


5

Example Transaction


6

Transaction Support

Can have one of two outcomes:– Success - transaction commits and database reaches a

new consistent state.

– Failure - transaction aborts, and database must be restored to consistent state before it started.

– Such a transaction is rolled back or undone.

Committed transaction cannot be aborted. Aborted transaction that is rolled back can be

restarted later.


7

State Transition Diagram for Transaction


8

Properties of Transactions

Four basic (ACID) properties of a transaction are:

Atomicity ‘All or nothing’ property.

Consistency Must transform database from one consistent state to another.

Isolation Partial effects of incomplete transactions should not be visible to other transactions.

Durability Effects of a committed transaction are permanent and must not be lost because of later failure.


9

DBMS Transaction Subsystem


10

Concurrency Control

Process of managing simultaneous operations on the database without having them interfere with one another.

Prevents interference when two or more users are accessing database simultaneously and at least one is updating data.

Although two transactions may be correct in themselves, interleaving of operations may produce an incorrect result.


11

Need for Concurrency Control

Three examples of potential problems caused by concurrency: – Lost update problem.– Uncommitted dependency problem.– Inconsistent analysis problem.


12

Lost Update Problem

Successfully completed update is overridden by another user.

T1 withdrawing £10 from an account with balx, initially £100.

T2 depositing £100 into same account. Serially, final balance would be £190.


13

Lost Update Problem

Loss of T2’s update avoided by preventing T1 from reading balx until after update.


14

Uncommitted Dependency Problem

Occurs when one transaction can see intermediate results of another transaction before it has committed.

T4 updates balx to £200 but it aborts, so balx should be back at original value of £100.

T3 has read new value of balx (£200) and uses value as basis of £10 reduction, giving a new balance of £190, instead of £90.


15

Uncommitted Dependency Problem

Problem avoided by preventing T3 from reading balx until after T4 commits or aborts.


16

Inconsistent Analysis Problem

Occurs when transaction reads several values but second transaction updates some of them during execution of first.

Sometimes referred to as dirty read or unrepeatable read.

T6 is totaling balances of account x (£100), account y (£50), and account z (£25).

Meantime, T5 has transferred £10 from balx to balz, so T6 now has wrong result (£10 too high).


17

Inconsistent Analysis Problem

Problem avoided by preventing T6 from reading balx and balz until after T5 completed updates.


18

Serializability

Objective of a concurrency control protocol is to schedule transactions in such a way as to avoid any interference.

Could run transactions serially, but this limits degree of concurrency or parallelism in system.

Serializability identifies those executions of transactions guaranteed to ensure consistency.


19

Serializability

ScheduleSequence of reads/writes by set of concurrent transactions.

Serial ScheduleSchedule where operations of each transaction are executed consecutively without any interleaved operations from other transactions.

No guarantee that results of all serial executions of a given set of transactions will be identical.


20

Nonserial Schedule

Schedule where operations from set of concurrent transactions are interleaved.

Objective of serializability is to find nonserial schedules that allow transactions to execute concurrently without interfering with one another.

In other words, want to find nonserial schedules that are equivalent to some serial schedule. Such a schedule is called serializable.


21

Serializability

In serializability, ordering of read/writes is important:

(a) If two transactions only read a data item, they do not conflict and order is not important.

(b) If two transactions either read or write completely separate data items, they do not conflict and order is not important.

(c) If one transaction writes a data item and another reads or writes same data item, order of execution is important.


22

Example of Conflict Serializability


23

Serializability

Conflict serializable schedule orders any conflicting operations in same way as some serial execution.

Under constrained write rule (transaction updates data item based on its old value, which is first read), use precedence graph to test for serializability.


24

Precedence Graph

Create:– node for each transaction;

– a directed edge Ti Tj, if Tj reads the value of an item written by TI;

– a directed edge Ti Tj, if Tj writes a value into an item after it has been read by Ti.

If precedence graph contains cycle schedule is not conflict serializable.


25

Example - Non-conflict serializable schedule

T9 is transferring £100 from one account with balance balx to another account with balance baly.

T10 is increasing balance of these two accounts by 10%.

Precedence graph has a cycle and so is not serializable.


26

Example - Non-conflict serializable schedule


27

View Serializability

Offers less stringent definition of schedule equivalence than conflict serializability.

Two schedules S1 and S2 are view equivalent if:– For each data item x, if Ti reads initial value of x in S1, Ti

must also read initial value of x in S2.– For each read on x by Ti in S1, if value read by x is

written by Tj, Ti must also read value of x produced by Tj in S2.

– For each data item x, if last write on x performed by Ti in S1, same transaction must perform final write on x in S2.


28

View Serializability

Schedule is view serializable if it is view equivalent to a serial schedule.

Every conflict serializable schedule is view serializable, although converse is not true.

It can be shown that any view serializable schedule that is not conflict serializable contains one or more blind writes.

In general, testing whether schedule is serializable is NP-complete.


29

Example - View Serializable schedule


30

Recoverability

Serializability identifies schedules that maintain database consistency, assuming no transaction fails.

Could also examine recoverability of transactions within schedule.

If transaction fails, atomicity requires effects of transaction to be undone.

Durability states that once transaction commits, its changes cannot be undone (without running another, compensating, transaction).


31

Recoverable Schedule

A schedule where, for each pair of transactions Ti and Tj, if Tj reads a data item previously written by Ti, then the commit operation of Ti precedes the commit operation of Tj.


32

Concurrency Control Techniques

Two basic concurrency control techniques:– Locking,– Timestamping.

Both are conservative approaches: delay transactions in case they conflict with other transactions.

Optimistic methods assume conflict is rare and only check for conflicts at commit.


33

Locking

Transaction uses locks to deny access to other transactions and so prevent incorrect updates.

Most widely used approach to ensure serializability. Generally, a transaction must claim a shared (read)

or exclusive (write) lock on a data item before read or write.

Lock prevents another transaction from modifying item or even reading it, in the case of a write lock.


34

Locking - Basic Rules

If transaction has shared lock on item, can read but not update item.

If transaction has exclusive lock on item, can both read and update item.

Reads cannot conflict, so more than one transaction can hold shared locks simultaneously on same item.

Exclusive lock gives transaction exclusive access to that item.


35

Locking - Basic Rules

Some systems allow transaction to upgrade read lock to an exclusive lock, or downgrade exclusive lock to a shared lock.


36

Example - Incorrect Locking Schedule

For two transactions above, a valid schedule using these rules is:

S = {write_lock(T9, balx), read(T9, balx), write(T9, balx), unlock(T9, balx), write_lock(T10, balx), read(T10, balx), write(T10, balx), unlock(T10, balx), write_lock(T10, baly), read(T10, baly), write(T10, baly), unlock(T10, baly), commit(T10), write_lock(T9, baly), read(T9, baly), write(T9, baly), unlock(T9, baly), commit(T9) }


37


If at start, balx = 100, baly = 400, result should be:

– balx = 220, baly = 330, if T9 executes before T10, or

– balx = 210, baly = 340, if T10 executes before T9.

However, result gives balx = 220 and baly = 340.

S is not a serializable schedule.


38


Problem is that transactions release locks too soon, resulting in loss of total isolation and atomicity.

To guarantee serializability, need an additional protocol concerning the positioning of lock and unlock operations in every transaction.


39

Two-Phase Locking (2PL)

Transaction follows 2PL protocol if all locking operations precede first unlock operation in the transaction.

Two phases for transaction:– Growing phase - acquires all locks but

cannot release any locks.– Shrinking phase - releases locks but cannot

acquire any new locks.


40

Preventing Lost Update Problem using 2PL


41

Preventing Uncommitted Dependency Problem using 2PL


42

Preventing Inconsistent Analysis Problem using 2PL


43

Cascading Rollback

If every transaction in a schedule follows 2PL, schedule is serializable.

However, problems can occur with interpretation of when locks can be released.


44

Cascading Rollback


45

Cascading Rollback

Transactions conform to 2PL. T14 aborts.

Since T15 is dependent on T14, T15 must also be rolled back. Since T16 is dependent on T15, it too must be rolled back.

This is called cascading rollback. To prevent this with 2PL, leave release of all

locks until end of transaction.


46

Concurrency Control with Index Structures

Could treat each page of index as a data item and apply 2PL.

However, as indexes will be frequently accessed, particularly higher levels, this may lead to high lock contention.

Can make two observations about index traversal:– Search path starts from root and moves down to leaf

nodes but search never moves back up tree. Thus, once a lower-level node has been accessed, higher-level nodes in that path will not be used again.


47


– When new index value (key and pointer) is being inserted into a leaf node, then if node is not full, insertion will not cause changes to higher-level nodes.

Suggests only have to exclusively lock leaf node in such a case, and only exclusively lock higher-level nodes if node is full and has to be split.


48


Thus, can derive following locking strategy:– For searches, obtain shared locks on nodes starting at root

and proceeding downwards along required path. Release lock on node once lock has been obtained on the child node.

– For insertions, conservative approach would be to obtain exclusive locks on all nodes as we descend tree to the leaf node to be modified.

– For more optimistic approach, obtain shared locks on all nodes as we descend to leaf node to be modified, where obtain exclusive lock. If leaf node has to split, upgrade shared lock on parent to exclusive lock. If this node also has to split, continue to upgrade locks at next higher level.


49

Deadlock

An impasse that may result when two (or more) transactions are each waiting for locks held by the other to be released.


50

Deadlock

Only one way to break deadlock: abort one or more of the transactions.

Deadlock should be transparent to user, so DBMS should restart transaction(s).

Three general techniques for handling deadlock: – Timeouts.– Deadlock prevention.– Deadlock detection and recovery.


51

Timeouts

Transaction that requests lock will only wait for a system-defined period of time.

If lock has not been granted within this period, lock request times out.

In this case, DBMS assumes transaction may be deadlocked, even though it may not be, and it aborts and automatically restarts the transaction.


52

Deadlock Prevention

DBMS looks ahead to see if transaction would cause deadlock and never allows deadlock to occur.

Could order transactions using transaction timestamps:– Wait-Die - only an older transaction can

wait for younger one, otherwise transaction is aborted (dies) and restarted with same timestamp.


53

Deadlock Prevention

– Wound-Wait - only a younger transaction can wait for an older one. If older transaction requests lock held by younger one, younger one is aborted (wounded).


54

Deadlock Detection and Recovery

DBMS allows deadlock to occur but recognizes it and breaks it.

Usually handled by construction of wait-for graph (WFG) showing transaction dependencies:– Create a node for each transaction.– Create edge Ti -> Tj, if Ti waiting to lock item

locked by Tj. Deadlock exists if and only if WFG contains

cycle. WFG is created at regular intervals.


55

Example - Wait-For-Graph (WFG)


56

Recovery from Deadlock Detection

Several issues:– choice of deadlock victim;– how far to roll a transaction back;– avoiding starvation.


57

Timestamping

Transactions ordered globally so that older transactions, transactions with smaller timestamps, get priority in the event of conflict.

Conflict is resolved by rolling back and restarting transaction.

No locks so no deadlock.


58

Timestamping

Timestamp

A unique identifier created by DBMS that indicates relative starting time of a transaction.

Can be generated by using system clock at time transaction started, or by incrementing a logical counter every time a new transaction starts.


59

Timestamping

Read/write proceeds only if last update on that data item was carried out by an older transaction.

Otherwise, transaction requesting read/write is restarted and given a new timestamp.

Also timestamps for data items:– read-timestamp - timestamp of last transaction

to read item;– write-timestamp - timestamp of last

transaction to write item.


60

Timestamping - Read(x)

Consider a transaction T with timestamp ts(T):

ts(T) < write_timestamp(x)

– x already updated by younger (later) transaction.

– Transaction must be aborted and restarted with a new timestamp.


61

Timestamping - Read(x)

ts(T) < read_timestamp(x)

– x already read by younger transaction.– Roll back transaction and restart it using a

later timestamp.


62

Timestamping - Write(x)

ts(T) < write_timestamp(x)

– x already written by younger transaction.– Write can safely be ignored - ignore obsolete

write rule.

Otherwise, operation is accepted and executed.


63

Example – Basic Timestamp Ordering


64

Comparison of Methods


65

Multiversion Timestamp Ordering

Versioning of data can be used to increase concurrency.

Basic timestamp ordering protocol assumes only one version of data item exists, and so only one transaction can access data item at a time.

Can allow multiple transactions to read and write different versions of same data item, and ensure each transaction sees consistent set of versions for all data items it accesses.


66

Multiversion Timestamp Ordering

In multiversion concurrency control, each write operation creates new version of data item while retaining old version.

When transaction attempts to read data item, system selects one version that ensures serializability.

Versions can be deleted once they are no longer required.


67

Optimistic Techniques

Based on assumption that conflict is rare and more efficient to let transactions proceed without delays to ensure serializability.

At commit, check is made to determine whether conflict has occurred.

If there is a conflict, transaction must be rolled back and restarted.

Potentially allows greater concurrency than traditional protocols.


68

Optimistic Techniques

Three phases:

– Read– Validation– Write


69

Optimistic Techniques - Read Phase

Extends from start until immediately before commit.

Transaction reads values from database and stores them in local variables. Updates are applied to a local copy of the data.


70

Optimistic Techniques - Validation Phase

Follows the read phase. For read-only transaction, checks that data read

are still current values. If no interference, transaction is committed, else aborted and restarted.

For update transaction, checks transaction leaves database in a consistent state, with serializability maintained.


71

Optimistic Techniques - Write Phase

Follows successful validation phase for update transactions.

Updates made to local copy are applied to the database.


72

Granularity of Data Items

Size of data items chosen as unit of protection by concurrency control protocol.

Ranging from coarse to fine:– The entire database.– A file.– A page (or area or database spaced).– A record.– A field value of a record.


73

Granularity of Data Items

Tradeoff: – coarser, the lower the degree of concurrency; – finer, more locking information that is needed

to be stored. Best item size depends on the types of

transactions.


74

Hierarchy of Granularity

Could represent granularity of locks in a hierarchical structure.

Root node represents entire database, level 1s represent files, etc.

When node is locked, all its descendants are also locked.

DBMS should check hierarchical path before granting lock.


75

Hierarchy of Granularity

Intention lock could be used to lock all ancestors of a locked node.

Intention locks can be read or write. Applied top-down, released bottom-up.


76

Levels of Locking


77

Database Recovery

Process of restoring database to a correct state in the event of a failure.

Need for Recovery Control– Two types of storage: volatile (main memory) and

nonvolatile.

– Volatile storage does not survive system crashes.

– Stable storage represents information that has been replicated in several nonvolatile storage media with independent failure modes.


78

Types of Failures

System crashes, resulting in loss of main memory.

Media failures, resulting in loss of parts of secondary storage.

Application software errors. Natural physical disasters. Carelessness or unintentional destruction of

data or facilities. Sabotage.


79

Transactions and Recovery

Transactions represent basic unit of recovery. Recovery manager responsible for atomicity and

durability. If failure occurs between commit and database

buffers being flushed to secondary storage then, to ensure durability, recovery manager has to redo (rollforward) transaction’s updates.


80

Transactions and Recovery

If transaction had not committed at failure time, recovery manager has to undo (rollback) any effects of that transaction for atomicity.

Partial undo - only one transaction has to be undone.

Global undo - all transactions have to be undone.


81

Example

DBMS starts at time t0, but fails at time tf. Assume data for transactions T2 and T3 have been written to secondary storage.

T1 and T6 have to be undone. In absence of any other information, recovery manager has to redo T2, T3, T4, and T5.


82

Recovery Facilities

DBMS should provide following facilities to assist with recovery:

– Backup mechanism, which makes periodic backup copies of database.

– Logging facilities, which keep track of current state of transactions and database changes.

– Checkpoint facility, which enables updates to database in progress to be made permanent.

– Recovery manager, which allows DBMS to restore database to consistent state following a failure.


83

Log File

Contains information about all updates to database:– Transaction records.– Checkpoint records.

Often used for other purposes (for example, auditing).


84

Log File

Transaction records contain:– Transaction identifier.– Type of log record, (transaction start, insert,

update, delete, abort, commit).– Identifier of data item affected by database

action (insert, delete, and update operations).– Before-image of data item.– After-image of data item.– Log management information.


85

Sample Log File


86

Log File

Log file may be duplexed or triplexed. Log file sometimes split into two separate

random-access files. Potential bottleneck; critical in determining

overall performance.


87

Checkpointing

CheckpointPoint of synchronization between database and log file. All buffers are force-written to secondary storage.

Checkpoint record is created containing identifiers of all active transactions.

When failure occurs, redo all transactions that committed since the checkpoint and undo all transactions active at time of crash.


88

Checkpointing

In previous example, with checkpoint at time tc, changes made by T2 and T3 have been written to secondary storage.

Thus:

– only redo T4 and T5,

– undo transactions T1 and T6.


89

Recovery Techniques

If database has been damaged:– Need to restore last backup copy of database and

reapply updates of committed transactions using log file.

If database is only inconsistent:– Need to undo changes that caused inconsistency.

May also need to redo some transactions to ensure updates reach secondary storage.

– Do not need backup, but can restore database using before- and after-images in the log file.


90

Main Recovery Techniques

Three main recovery techniques:

– Deferred Update– Immediate Update– Shadow Paging


91

Deferred Update

Updates are not written to the database until after a transaction has reached its commit point.

If transaction fails before commit, it will not have modified database and so no undoing of changes required.

May be necessary to redo updates of committed transactions as their effect may not have reached database.


92

Immediate Update

Updates are applied to database as they occur. Need to redo updates of committed transactions

following a failure. May need to undo effects of transactions that

had not committed at time of failure. Essential that log records are written before

write to database. Write-ahead log protocol.


93

Immediate Update

If no “transaction commit” record in log, then that transaction was active at failure and must be undone.

Undo operations are performed in reverse order in which they were written to log.


94

Shadow Paging

Maintain two page tables during life of a transaction: current page and shadow page table.

When transaction starts, two pages are the same. Shadow page table is never changed thereafter

and is used to restore database in event of failure. During transaction, current page table records

all updates to database. When transaction completes, current page table

becomes shadow page table.


95

Advanced Transaction Models

Protocols considered so far are suitable for types of transactions that arise in traditional business applications, characterized by:– Data has many types, each with small number

of instances.– Designs may be very large.– Design is not static but evolves through time. – Updates are far-reaching.– Cooperative engineering.


96


May result in transactions of long duration, giving rise to following problems:– More susceptible to failure - need to minimize

amount of work lost.– May access large number of data items -

concurrency limited if data inaccessible for long periods.

– Deadlock more likely.– Cooperation through use of shared data items

restricted by traditional concurrency protocols.


97


Look at five advanced transaction models:

– Nested Transaction Model– Sagas– Multi-level Transaction Model– Dynamic Restructuring– Workflow Models


98

Nested Transaction Model

Transaction viewed as hierarchy of subtransactions. Top-level transaction can have number of child

transactions. Each child can also have nested transactions. In Moss’s proposal, only leaf-level subtransactions

allowed to perform database operations. Transactions have to commit from bottom upwards. However, transaction abort at one level does not have

to affect transaction in progress at higher level.


99


Parent allowed to perform its own recovery:– Retry subtransaction.– Ignore failure, in which case subtransaction

non-vital.– Run contingency subtransaction.– Abort.

Updates of committed subtransactions at intermediate levels are visible only within scope of their immediate parents.


100


Further, commit of subtransaction is conditionally subject to commit or abort of its superiors.

Using this model, top-level transactions conform to traditional ACID properties of flat transaction.


101

Example of Nested Transactions


102

Nested Transaction Model - Advantages

Modularity - transaction can be decomposed into number of subtransactions for purposes of concurrency and recovery.

Finer level of granularity for concurrency control and recovery.

Intra-transaction parallelism. Intra-transaction recovery control.


103

Emulating Nested Transactions using Savepoints

An identifiable point in flat transaction representing some partially consistent state.

Can be used as restart point for transaction if subsequent problem detected.

During execution of transaction, user can establish savepoint, which user can use to roll transaction back to.

Unlike nested transactions, savepoints do not support any form of intra-transaction parallelism.


104

Sagas

“A sequence of (flat) transactions that can be interleaved with other transactions”.

DBMS guarantees that either all transactions in saga are successfully completed or compensating transactions are run to undo partial execution.

Saga has only one level of nesting. For every subtransaction defined, there is

corresponding compensating transaction that will semantically undo subtransaction’s effect.


105

Sagas

Relax property of isolation by allowing saga to reveal its partial results to other concurrently executing transactions before it completes.

Useful when subtransactions are relatively independent and compensating transactions can be produced.

May be difficult sometimes to define compensating transaction in advance, and DBMS may need to interact with user to determine compensation.


106

Multi-level Transaction Model

Closed nested transaction - atomicity enforced at the top-level.

Open nested transactions - allow partial results of subtransactions to be seen outside transaction.

Saga model is example of open nested transaction. So is multi-level transaction model where tree of

subtransactions is balanced. Nodes at same depth of tree correspond to

operations of same level of abstraction in DBMS.


107

Multi-level Transaction Model

Edges represent implementation of an operation by sequence of operations at next lower level.

Traditional flat transaction ensures no conflicts at lowest level (L0).

In multi-level model two operations at level Li may not conflict even though their implementations at next lower level Li-1 do.


108

Example - Multi-level Transaction Model


109

Example - Multi-level Transaction Model

T7: T71, which increases balx by 5

T72, which subtracts 5 from baly

T8: T81, which increases baly by 10

T82, which subtracts 2 from balx

As addition and subtraction commute, can execute these subtransactions in any order, and correct result will always be generated.


110

Dynamic Restructuring

To address constraints imposed by ACID properties of flat transactions, two new operations proposed: split_transaction and join_transaction.

split-transaction - splits transaction into two serializable transactions and divides its actions and resources (for example, locked data items) between new transactions.

Resulting transactions proceed independently.


111


Allows partial results of transaction to be shared, while still preserving its semantics.

Can be applied only when it is possible to generate two transactions that are serializable with each other and with all other concurrently executing transactions.


112


Conditions that permit transaction to be split into A and B are:– .AWriteSet BWriteSet BWriteLast.

If both A and B write to same object, B’s write operations must follow A’s write operations.

– .AReadSet BWriteSet = . A cannot see any results from B.

– .BReadSet AWriteSet = ShareSet. B may see results of A.


113


These guarantee that A is serialized before B. However, if A aborts, B must also abort. If both BWriteLast and ShareSet are empty,

then A and B can be serialized in any order and both can be committed independently.


114


join-transaction - performs reverse operation, merging ongoing work of two or more independent transactions, as though they had always been single transaction.


115


Main advantages of dynamic restructuring are:

Adaptive recovery. Reducing isolation.


116

Workflow Models

Has been argued that above models are still not powerful to model some business activities.

More complex models have been proposed that are combinations of open and nested transactions.

However, as they hardly conform to any of ACID properties, called workflow model used instead.

Workflow is activity involving coordinated execution of multiple tasks performed by different processing entities (people or software systems).


117

Workflow Models

Two general problems involved in workflow systems: – specification of the workflow, – execution of the workflow.

Both problems complicated by fact that many organizations use multiple, independently managed systems to automate different parts of the process.


1 chapter 20 transaction management transparencies © pearson education limited 1995, 2005

Documents

transaction support

u aborted transaction

u transforms database

u properties of transactions

properties of transactions

u concurrency control

concurrency control

committed transaction